Faculty of Science

Permanent URI for this communityhttp://hdl.handle.net/20.500.12280/177

Browse

Browsing Faculty of Science by Subject "AI governance"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Artificial intelligence as a competitive equaliser for mid-sized retailers
    (SSRN, Elsevier Ltd., 2026-08-28) Bazigu, Rodgers
    Mid-sized retailers increasingly compete against e-commerce platforms that offer large-scale personalisation, predictive supply chains, and adaptive marketing at a cost smaller firms cannot easily match. This paper develops an artificial intelligence (AI) adoption framework for a representative midsized retailer, referred to here as StyleHub, addressing four linked questions: which AI applications can meaningfully address a mid-sized retailer's core operational challenges; which design factors govern the construction of an effective personalisation model; how personalisation can be pursued without eroding customer trust; and what a realistic, phased adoption pathway looks like. Drawing on the retail AI, marketing ethics, and AI governance literatures, the paper argues that competitive advantage at StyleHub's scale depends less on acquiring the most advanced tools available and more on sequencing investment around organisational readiness, data quality, and safeguards built in from the outset rather than retrofitted later. Two contrasting adoption trajectories are outlined, one where technical investment outpaces governance and stalls, and one where phased, ethically grounded adoption compounds gradually, to argue that responsible implementation is itself a source of durable differentiation for resource-constrained retailers.
  • Thumbnail Image
    Item
    From data to decisions: a staged framework for AI, NLP, and computer vision adoption at InnovateCorp
    (SSRN, Elsevier Ltd., 2026-08-17) Bazigu, Rodgers
    Firms accumulating transactional, sensor, textual, and visual data frequently lack the analytical infrastructure to convert that data into decisions, a gap this paper examines through the case of a representative technology-enabled enterprise, InnovateCorp. The analysis develops a staged framework spanning four domains: predictive and prescriptive analytics for strategic decision-making; neural network architectures for predictive maintenance, inventory, and scheduling; natural language processing and conversational AI for customer engagement; and computer vision for quality inspection and retail intelligence. Rather than treating these as independent technology initiatives, the paper argues that they form concentric layers of a single organisational capability, one built on a unified data warehouse, staged deployment, and continuous human oversight rather than on the sophistication of any individual model. Reviewing evidence on ensemble methods, recurrent and convolutional architectures, transformer-based language models, and detection networks, the paper finds that the dominant constraint on value realisation is organisational rather than algorithmic: governance, data quality, and employee adoption determine outcomes considerably more than model selection. The paper concludes that mainstream, well-understood architectures, deployed conservatively and in sequence, offer InnovateCorp a more defensible path to sustained analytical advantage than pursuing technically novel but organisationally unproven alternatives.